AMBER-ii Force Field

AMBER-ii is an extension of AMBER03 one of the most successful force fields for simulating proteins. The idea is keeping as far as possible AMBER03 parameters for protein to expand it to other classes of substances and to correct known problems.

Hydrocarbons

It is known that AMBER and OPLS parametrizations describe well low-molecular alkanes, but are not suitable for work with high-molecular alkanes. Consequently, AMBER is little suited to lipid modeling. Later, a series of lipid parametrizations compatible with AMBER was developed. AMBER-ii is one such parameterization. It well describes both the conformational and thermodynamic properties of alkanes, regardless of their length. In addition, this parametrization is compatible with parametrization for perfluoroalkanes.

Perfluoroalkanes

The main problem perfluoroalkanes force fields is their compatibility with the force fields for alkanes. It is experimentally known that liquid perfluoroalkanes are poorly mixed with conventional oils. This property, weakly interact with both water and oils, is widely used to create non-stick coatings (Teflon). However, this property is difficult to reproduce in simulations. AMBER-ii reproduces this with good accuracy.

If accurate alkane force fields are important for modeling natural lipids, then perfluoroalkane force fields open the possibility of modeling artificial fluoride-containing lipids. Artificial fluorinated lipids provide interesting possibilities for creating self-assembling spatial structures. For example, if natural lipids form bilayers, the addition of fluorinated agents can result in the formation of a trilayer.

Heavy elements

AMBER parameterizations use the Lorentz−Berthelot combining rules. This works great for elements of the second period. Parametrization of the third period elements is not so good. Nevertheless, for sulfur and phosphorus, AMBER provides acceptable parameters. Transition to the elements of the fourth and subsequent periods is hardly possible on the basis of the Lorentz−Berthelot rules. The error in the dispersion interactions can reach two times.

Several variants of the rules that solve this problem are suggested, however, they give the worst results for hydrogen. Since hydrogen is universally present in organic substances, the idea of transition to these rules is unattractive. In AMBER-ii, hydrogen is considered as a special element. Interactions with it are described by the Lorentz-Berdelot rules. For all other elements, the Waldman-Hagler rules are adopted. This choice preserves the majority of AMBER parameters for proteins. Only sulfur has to be re-parametrized. Which was done in OPLS style, using experimental data on density and heat of evaporation of liquids. The resultant force field yields for the interactions in proteins results almost indistinguishable from the original AMBER03.

Metal ions

In this case, AMBER-ii departs from the original parameterization. First of all, it should be noted that AMBER does not have its own ion parametrization. Parametrization was borrowed from the force field of other authors. This parametrization was developed for the description of ions in liquid water. This state is characterized by high coordination numbers. In proteins low coordination numbers are usual for metals. As a result, the interaction of ions, especially multiply charged ones, is poorly modeled.

Following the arguments of the Leontiev and Stuchebryukhov in AMBER-ii the scheme with decreasing integer charges on the ions is accepted. This approach makes it possible to develop parametrizations for both single-charged and multicharged ions in simple algorithmic manner. The resulting force field yields acceptable results both in the free energy of hydration of ions in water and in coordination numbers in proteins.

It was demonstrated that with the resulting parametrization protein folding can be achieved, both for alpha and beta proteins.